A mouse scroll test on Mac records the raw browser wheel events delivered while the pointer sits over one visible test area, then classifies each one as up, down, left, right, or none and keeps signed cumulative X and Y totals for the current tab. The numbers you see are exactly the deltaX and deltaY values that your Mac's browser chose to deliver, nothing more and nothing less. That makes the test a quick way to confirm whether a wheel or trackpad gesture reaches the page at all, in which direction it landed, and whether the magnitudes look stable across repeated moves. It does not measure physical centimeters, USB polling rate, hardware wear, or any system setting outside the browser. To get useful answers on macOS, you also need to know how Safari and other Mac browsers package trackpad and Magic Mouse gestures, because the same gesture can arrive as very different delta numbers depending on acceleration curves and natural-scrolling state.
What a Mac scroll test actually captures
The Mouse Scroll Test draws a dashed target zone on the page and starts listening for wheel events as soon as your pointer enters it. Everything outside that zone is ignored, so the rest of the page can scroll normally when you move over it but stays still when you scroll inside the dashed area. Each event is recorded as four pieces of raw data: the total event count, a directional bucket, the latest horizontal and vertical delta pair, and the signed running totals for X and Y.
The directional bucket is a single label per event and follows a fixed rule. If both deltaX and deltaY arrive as zero, the event is filed under none, which preserves silent or synthetic events instead of dropping them. Otherwise the tool compares the absolute magnitudes, lets vertical win ties, and reads the sign of the dominant axis to assign up, down, left, or right. The signed cumulative totals can grow positive or negative as you scroll in either direction, and opposite movements can cancel each other out, so a near-zero cumulative total does not mean no scrolling happened.
macOS settings that change what the test shows
macOS exposes two scroll-related preferences that change the direction labels you will see in the test. The first is the natural scrolling toggle in System Settings, under Trackpad and Mouse, which reverses the direction the page moves when you roll the wheel or swipe on a trackpad. With natural scrolling on, a wheel rolled toward you is delivered to the browser as a negative deltaY, which the test labels down, so the page moves down even though the wheel turned toward you. Turn it off and the same wheel motion arrives as a positive deltaY, which the test labels up, while the page moves up. Neither outcome is a bug, and both are simply the same physical gesture routed through different signed deltas.
The second preference is scroll speed or acceleration, which on macOS scales the magnitude of each delta without changing its sign. A fast trackpad swipe can arrive as a deltaY of around 12, while a slow swipe on the same trackpad arrives as a deltaY of around 2, even though both events point the same direction. The test will simply show two different numbers for two different gestures, which is the right answer. If you want comparable numbers, slow your gestures down, use a fixed hand motion, and repeat the same gesture several times.
How to run a mouse scroll test on Mac
Open the Mouse Scroll Test page in the browser you normally use on your Mac, whether that is Safari, Chrome, Firefox, or Arc. Move the pointer into the dashed test area, then follow these steps in order.
- Place the pointer over the dashed test area and move the wheel or trackpad vertically and horizontally.
- Compare total events, directional buckets, the latest delta pair, and signed cumulative X and Y values.
- Repeat under the same browser and device settings, then select Reset test to clear the local session.
For a clean first pass, use four short gestures in sequence: one slow roll away from you, one slow roll toward you, one left swipe on a Magic Mouse or trackpad, and one right swipe. Then scroll continuously for about five seconds in each direction at a steady pace to see whether the deltas stay roughly stable or whether they spike. If you have more than one pointing device on the Mac, such as a built-in trackpad and a USB mouse, switch to each one and run the same four gestures to compare. Once you finish, click Reset test to clear the count and start fresh for a second run under identical conditions.
Reading the directional buckets, latest delta, and cumulative sums
The numbers in the test follow the same browser rules on macOS as on any other platform, so reading them on a Mac is mostly about knowing which axis you are looking at and what each label means. The table below summarizes the classification rule, which the tool applies to every event delivered inside the dashed area.
| Axis comparison | Direction bucket |
|---|---|
| Both deltas are 0 | none |
| |vertical| greater than or equal to |horizontal|, vertical nonzero | up (positive deltaY) or down (negative deltaY) |
| |horizontal| greater than |vertical|, horizontal nonzero | right (positive deltaX) or left (negative deltaX) |
| |vertical| equals |horizontal|, both nonzero | up or down (vertical wins the tie) |
When you read the latest pair, match the dominant axis label to the gesture you just made. If you rolled the wheel away from you and saw a positive deltaY labeled up, the test is receiving events correctly for your natural-scrolling setting. The signed cumulative totals are most useful for spotting drift: if cumulative X climbs steadily in one direction while you alternate left and right scrolls, the magnitudes are not balanced and one gesture is producing bigger raw deltas than the other. Use the directional counts and the total event count as the main record of activity, because cumulative sums can cancel each other out.
If the numbers still feel unclear, the Mouse Scroll Test accuracy guide walks through why raw deltas should not be treated as calibrated distance.
Why raw deltas differ between Macs and browsers
A Mac with a Magic Mouse, a MacBook with a built-in trackpad, and a Mac connected to a third-party USB wheel mouse will each deliver different delta values for what feels like the same gesture, and that is by design. The browser only sees the WheelEvent the operating system and driver handed to it, and those layers can interpret one notch of scroll as 1, 3, 4, or any other number depending on acceleration, accessibility zoom, and driver choice. macOS itself changes scroll behavior between System Settings versions, so the same Mac on one major macOS release and on another can produce different delta values for the same finger motion.
Browsers add their own variation. Safari and Chrome both smooth and combine high-frequency trackpad events before they reach the page, Firefox tends to deliver smaller and more granular deltas, and Arc behaves closer to Chrome on the same engine. All of that is browser-defined event data, not physical distance. The test deliberately does not normalize the values to pixels per scroll, lines per notch, centimeters of finger travel, or any cross-device score. A test on a Mac is therefore most informative when you compare two runs on the same Mac, in the same browser, with the same System Settings, rather than across machines.
When the test points to something else
A missing direction in the test does not always mean a broken mouse. Some pointing devices simply cannot produce that axis: a traditional vertical wheel will not deliver horizontal deltas no matter how hard you try, and a basic trackpad may not produce left and right wheel events even though it supports horizontal swipes for the OS. The browser can also suppress wheel events, smooth them past the point of recognition, or hand them to a gesture system before the page sees them. Remote desktop software, accessibility zoom, and third-party scrolling utilities sit between macOS and the browser and can drop or rewrite events on the way through.
Treat the test as proof of what reached this page during this session, in this tab, and nothing more. If a direction fails only in a specific app, test that app separately. If a direction fails only when Bluetooth is enabled, pair the device again or check radio interference. If hardware health or repair decisions depend on the result, use Apple Diagnostics, a Magic Mouse utility, or qualified support instead of a browser-based wheel test. The tool is honest about its narrow contract: it counts and classifies the finite browser wheel deltas that arrive inside its visible zone, and it leaves hardware diagnostics to the appropriate tool.
For a deeper look, see Mouse Button Test: Decode What Your Browser Sees.